linux-kernel - Re: [REGRESSION 4.11] Commit d8514d8edb5b ("ovl: copy up regular file using O

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20170330092831.6c88f362@gmail.com>
Date:   Thu, 30 Mar 2017 09:28:31 +0200
From:   Ralph Sennhauser <ralph.sennhauser@...il.com>
To:     Richard Weinberger <richard@....at>
Cc:     Amir Goldstein <amir73il@...il.com>,
        Miklos Szeredi <miklos@...redi.hu>,
        linux-unionfs@...r.kernel.org,
        linux-kernel <linux-kernel@...r.kernel.org>,
        linux-mtd@...ts.infradead.org, regressions@...mhuis.info,
        Artem Bityutskiy <dedekind1@...il.com>,
        Adrian Hunter <adrian.hunter@...el.com>,
        David Oberhollenzer <goliath@...ma-star.at>
Subject: Re: [REGRESSION 4.11] Commit d8514d8edb5b ("ovl: copy up regular
 file using O_TMPFILE") breaks ubifs

Hi Richard,

On Thu, 30 Mar 2017 08:56:57 +0200
Richard Weinberger <richard@....at> wrote:

>     
> Ralph,
> Please apply this
> fix:http://lists.infradead.org/pipermail/linux-mtd/2017-March/072613.html
> 

No longer able to observe an issue with this one on top. The two
patches seem all that was needed to get things back to a working state.

Do you want me to test anything else than linux-next?

Thanks
Ralph

> 
> Von meinem Samsung Gerät gesendet.
> 
> -------- Ursprüngliche Nachricht --------
> Von: Ralph Sennhauser <ralph.sennhauser@...il.com> 
> Datum: 30.03.17  07:53  (GMT+01:00) 
> An: Richard Weinberger <richard@....at> 
> Cc: Amir Goldstein <amir73il@...il.com>, Miklos Szeredi
> <miklos@...redi.hu>, linux-unionfs@...r.kernel.org, linux-kernel
> <linux-kernel@...r.kernel.org>, linux-mtd@...ts.infradead.org,
> regressions@...mhuis.info, Artem Bityutskiy <dedekind1@...il.com>,
> Adrian Hunter <adrian.hunter@...el.com>, David Oberhollenzer
> <goliath@...ma-star.at> Betreff: Re: [REGRESSION 4.11] Commit
> d8514d8edb5b ("ovl: copy up regular&nbsp; file using O_TMPFILE") breaks ubifs 
> 
> Hi Richard,
> 
> On Thu, 30 Mar 2017 00:15:31 +0200
> Richard Weinberger <richard@....at> wrote:
> 
> > Ralph,
> > 
> > Am 29.03.2017 um 23:26 schrieb Ralph Sennhauser:  
> > >> # create and link a tmpfile - then remove it
> > >> sudo rm -rf foo; sudo xfs_io -T -c "flink foo" . ; ls -l foo;
> > >> sudo rm foo    
> > > 
> > > next-20170328, obviously without your patch:
> > > 
> > >   # xfs_io -T -c "flink foo" .
> > >   .: Not supported    
> > 
> > -T tells xfs_io to create a tmpfile but you have it disabled in
> > UBIFS.  
> 
> Bash history says I booted the other partition than I flashed (uname
> just happened to match), the downside of the dual boot layout. :)
> 
>   # rm -rf foo
>   # xfs_io -T -c "flink foo" .
>   # ls -l foo
>   -rw-------    1 root     root             0 Mar 30 02:00 foo
>   # rm foo
>   [  305.001436] UBIFS error (ubi0:0 pid 2493): ubifs_add_orphan:
> orphaned twice
> 
> However, unlike with the overlay setup the filesystem can be mounted
> just fine without imediatly visible side effects.
> 
> > 
> > Anyway, can you please test the attached patch?
> > Amir was right, UBIFS misses a corner case in ubifs_link(). ;-\
> > 
> > I think I understand also why we never noticed this in xfstests,
> > UBIFS prints only a non-fatal error while umounting the filesystem
> > in this case. But will test tomorrow in more detail, need some
> > sleep now.
> > 
> > diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c
> > index 30825d882aa9..95e348a2da29 100644
> > --- a/fs/ubifs/dir.c
> > +++ b/fs/ubifs/dir.c
> > @@ -748,6 +748,9 @@ static int ubifs_link(struct dentry *old_dentry,
> > struct inode *dir, goto out_fname;
> > 
> >  	lock_2_inodes(dir, inode);
> > +	if (inode->i_nlink == 0)
> > +		ubifs_delete_orphan(c, inode->i_ino);
> > +
> >  	inc_nlink(inode);
> >  	ihold(inode);
> >  	inode->i_ctime = ubifs_current_time(inode);
> > 
> > Thanks,
> > //richard  
> 
> With this patch I'm no longer able to reproduce _this_ issue, however,
> rename no longer works either:
> 
>   # mv /etc/config/wireless /etc/config/wireless.back
>   mv: can't rename '/etc/config/wireless': Invalid argument
>   # ls /etc/config/wireless
>   /etc/config/wireless
>   # rm /etc/config/wireless
>   # ls /etc/config/wireless
>   ls: /etc/config/wireless: No such file or directory
> 
> 
> Thanks
> Ralph